The Agent Skills Directory

[SAFE]: The skill provides documentation and educational content for AI developers. No malicious instructions, exfiltration, or obfuscation were found.
[PROMPT_INJECTION]: The skill contains common prompt injection phrases (e.g., 'Ignore all previous instructions', 'reveal your system prompt') used as educational examples in the 'Sharp Edges' section. These are identified as threats to mitigate, not instructions for the agent to follow.
[EXTERNAL_DOWNLOADS]: Code examples reference industry-standard packages from well-known services. These include 'zod', 'openai', 'tiktoken', and the Vercel AI SDK ('ai').
[CREDENTIALS_UNSAFE]: The skill actively encourages safe credential management by recommending environment variables and flagging hardcoded keys as security errors.
[DATA_EXFILTRATION]: No patterns of sensitive data access or unauthorized external transmission were detected. The network operations described in the code snippets are functional parts of the educational examples for calling established AI APIs.

ai-product